Separate the survey respondents into groups. You might be interested in only the results of surveys completed by people within a certain demographic, such as gender, race, religion or location. If this is the case, weed out the survey results in which you aren't interested.
Read all of the survey responses before you interpret the data. This is especially important for surveys that include open-ended questions that are more difficult that others to categorize. As you read, start to think about how you will separate your data into meaningful categories.
Create sets of data from the survey results. This involves separating the survey respondents into categories based on their answers. For example, all females who answered "no" to a question might be in one set, while those who answered "yes" would be in another set. Each survey respondent may be in several sets, depending on the types of questions you asked and how many questions are interpreted at one time.
Create percentages to help you interpret your survey results when applicable. For example, if 16 out of 100 respondents in your demographic group answered "no" to a question, then the percentage is a 16 percent negative response. Large sample sizes have more accurate percentages than small sample sizes.
Pay attention to questions that weren't answered at all or that weren't fully answered. These questions may have been poorly worded or too confusing for respondents, which affects the accurate interpretation of your survey. Also, if most respondents answered a question the same way, you could need to vary the demographics of your respondents, or the question could be flawed, too easy or too misleading, which also will skew your results.